WO2008028512A1

WO2008028512A1 - Bolometer and method for producing a bolometer

Info

Publication number: WO2008028512A1
Application number: PCT/EP2006/008790
Authority: WO
Inventors: Holger Vogt; Marco Russ
Original assignee: Fraunhofer-Gesellschaft zur Förderung der angewandten Forschung e.V.
Priority date: 2006-09-08
Filing date: 2006-09-08
Publication date: 2008-03-13
Also published as: CN101512308A; DE112006004013B4; DE112006004013A5; US8080797B2; US20090321644A1; CN101512308B

Abstract

A bolometer comprises a membrane, a first spacer and a second spacer, wherein the membrane comprises a resistance layer and a contact layer. On a side facing a substrate, the contact layer has a first contact region, in which the first spacer electrically contact-connects the contact layer, and a second contact region, in which the second spacer electrically contact-connects the contact layer. This keeps the membrane at a predetermined distance from the substrate. The contact layer is interrupted in the lateral direction by a gap, with the result that the contact layer is divided at least into two parts, the first part of which has the first contact region and the second part of which has the second contact region, and there is no direct connection between the first contact region and the second contact region within the contact layer, wherein the resistance layer is in contact with the first part of the contact layer and the second part of the contact layer.

Description

Bolometer and method of making a bolometer

description

The present invention relates to a bolometer and a method of making a bolometer, and more particularly to a scalable microbolometer structure.

A bolometer is a device for measuring an electromagnetic radiation intensity of a certain wavelength range (about 3-15 μm). It comprises an absorber, which converts the electromagnetic radiation into heat, and a device for measuring a temperature increase. Depending on the heat capacity of the material, there is a direct correlation between an amount of absorbed radiation and the resulting increase in temperature. Thus, the temperature increase can serve as a measure of an intensity of a sunken radiation. Particularly interesting are bolometers for measuring infrared radiation, where most bolometers have the highest sensitivity.

A bolometer may be used in the art as an infrared sensor, night vision imaging imager or thermal imaging camera.

A bolometer which serves as an infrared sensor comprises a thin layer, which is arranged thermally isolated in the sensor, eg suspended as a membrane. The infrared radiation is absorbed in this membrane, thereby increasing its temperature. If this membrane consists of a metallic or preferably a semiconducting material, then the electrical resistance changes depending on the temperature increase and the temperature coefficient of resistance of the material used. Examples of different materials are in the document: RA Wood: "Monolithic Silicon Microbolometer Arrays," Semiconductor Semimetals, vol. 47, pp. 43-121, 1997. alternative the membrane is an insulator (silicon oxide or silicon nitride) on which the resistor has been deposited as a further thin layer. In other embodiments, insulating layers and an absorber layer are provided in addition to the resistive layer.

The temperature dependence of metal film resistors is linear, and semiconductors as resistive material have an exponential dependence. Diodes as heat detectors with their current-voltage characteristic also promise high dependency

I ₀ = I ₀ * (Exp {eU _D / kT} -l)

where T is the temperature, k is the Boltzmann constant, e is the elementary electric charge, I ₀ and U _{D are} a current and voltage in the diode, and Io is a voltage-independent constant.

Bolometers can serve as individual sensors, but can also be constructed as rows or 2D arrays. Lines and arrays are typically produced today using microsystem technology in surface micromachining on a silicon substrate. This is called microbolometer arrays.

A preferred wavelength of the infrared radiation to be detected is 8-14 microns, since in this wavelength range radiates matter that has approximately room temperature (300 K). The wavelength range of 3-5 μm is also interesting because of a permeable atmospheric window.

A significant advantage of thermal bolometers over other (IR) photonic IR (IR) detectors is that they are at room temperature, i. can be operated without cooling.

The aim of further development is to arrange as many bolometer cells (pixels) as possible in an array. Thus, points The array has a higher pixel count and provides a better resolution of an image with the same total area (chip area) of the array. So an arrangement of 160 x 120 pixels is common, 320 x 240 also available, 640 x 480 pixels (VGA resolution) are announced, but only at considerable additional cost soon available. At the same time, it makes sense to minimize the cost of the array in order to open up new markets, such as the automotive sector.

The usual dimensions of a single pixel in the array include a pixel area of 35 x 35 μm ² to 50 x 50 μm ² . Thus, a chip area at 320 × 240 pixels is at least 12.2 × 8.4 mm ² = 94 mm ² (pure pixel area) plus an area for a readout circuit (eg 2 mm in addition per edge), in the sum of about 137 mm ² . Since a yield (the number of good chips based on the total number on a disk) decreases sharply with increasing chip area, economical production of such an array is hardly possible. Thus, increasing the number of pixels should be accompanied by a reduction in pixel area. IR imagers with 35 x 35 μm ² are now commercially available. As in the document http: // www-leti.cea. fr / commun / AR-2005 / pdf / t7-2-Mottin.pdf, there are 25 x 25 μm ² arrays under development. But even this already scaled surface leads to image receivers with VGA resolution to an unreasonably large chip area (estimated approximately 250 mm ² ). Further scaling of the pixel area is therefore absolutely necessary. The target is pixels with a pixel area of approx. 15 x 15 μm ² . A further reduction is countered by the fact that the optics used then would have to have a very high quality, which in turn would only be feasible at very high costs.

A detection of infrared radiation in the microbolometer is based on the fact that the radiation heats a thermally well insulated resistor. This is temperature dependent and therefore changes its resistance as a function of the heating. A resistance change is made via a ROIC (ROIC = read out i.ntegrated circuit) read out. Typical temperature increases at the resistor are in the range of less millikelvin (mK) per degree of temperature change of an observed target. In order for this temperature increase to be possible at the bolometer, the resistance must be thermally insulated very well. This is achieved by placing the resistor on a membrane (or forming it as a membrane itself), which is placed a few microns above a wafer surface and bonded to the wafer surface or substrate at only a few points of low thermal conductivity ,

Figure 5 shows two of a prior art bolometer disclosed in R.A. Wood: "Monolithic Silicon Microbolometer Arrays," Semiconductor Semimetals, vol. 47, pp. 43-121, 1997. FIG. 5 a shows a single-level pixel with a sensor 51, electronics 52, which is located on a substrate 54 and has a pixel size 53. FIG. 5b shows a two-level pixel where the electronics 52 are mounted below the sensor 51. This bolometer also corresponds to the state of the art and has a higher fill factor (ratio of IR-sensitive area to total area) compared to the bolometer shown in FIG. 5a.

The membrane is produced, for example, by making the resistor or sensor 51 on a wafer surface 55 and then under-etching the region so that a cavity 56 is created. By locally removing, for example, silicon (Si), the thermal resistance between resistor on the diaphragm 51 and the substrate 54 increases. A readout circuit 52 is integrated adjacent to the diaphragm 51 and thus costs additional chip area. Therefore, a structure of Fig. 5b is more advantageous in which the resistor 51 is mounted in a second plane on a membrane above the read-out circuit 52. To measure the resistance, two contact points are necessary. They can be formed by arranging feed lines on upwardly inclined sections of the membrane. The bevels also serve as spacers for the membrane. FIG. 6 shows a perspective view of a corresponding structure with a membrane 10 consisting of a carrier 35 and a resistance layer 18. Such an arrangement is described in Fig. 2 in US Patent 5,688,699 (Nov. 2, 1997; BT Cunningham, BI Patel: "Microbolometer"). The membrane 10 is held by oblique support arms 20 having an electrically conductive layer 32 and a thermal insulating layer 22. A contact of the membrane 10 via the support arms 20 has an overlap 33 and the support arms 20 extend into an epitaxial pit 14, where the corresponding circuit (not shown in the figure) is located. The epitaxial layer 14 is located between a substrate 12 and an insulating layer 24.

If the membrane 10 is flat (without bevels), the signal is supplied via metallic plugs, which at the same time serve as spacers. This structure is in http: // www-leti. cea. Fig. 7 shows a perspective view of such a conventional structure with a membrane 10 on two contact plugs 26a and 26b spaced at a distance 72 from one another Underground 73 is held. The membrane 10 having a size 75 has a thickness 74 and the substrate 73 comprises a reflector. A thermal insulation to the substrate 73 is thereby produced via the bridges 76a, b. On the substrate 73 is a ROIC Input Päd 77, via which the bolometer is contacted. The contacting of the membrane 10 in this case has an overlap 78 in comparison to a diameter of the contact plugs 26a and 26b. This overlap 78 reduces the fill factor.

Optimum absorption of the IR radiation is obtained by the membrane 10 corresponding to a sheet resistance a propagation resistance of an electromagnetic wave in air (377 Ω / D) and in the amount of λ / 4 (about 2.5 microns at the preferred wavelength λ, for example, 8-14 microns) is disposed over a reflector 73.

US Pat. No. 5,912,464 mentions such a bolometer and a method of manufacture, and FIG. 8 shows a section. FIG. 8 a shows a cross-section through a contacting of the membrane 10, the cross-sectional plane being shown by a dashed-dotted line in FIG. 8 b with a viewing direction 81.

The contact plug 26b contacts a connection pad 77 and at the same time a contact layer 23. Further layers of the bolometer are a reflection layer 21, a sacrificial layer 22, the bolometer or resistance layer 27 and transition layers 24 and 25. The electrical contacting of the resistance layer 27 is established via the contact layer 23 and the transition layers 24 and 25 serve for better contacting of the contact layer 23. The contact layer 23 extends in a serpentine fashion along the resistance layer 27 from a contact plug 26a to the contact plug 26b. The serpentine configuration of the electrode layer 23 is shown by a dashed line 82 in FIG. 8b. The serpentine configuration of the electrode layer 23 serves to better absorption of the infrared radiation.

Also in this bolometer according to the prior art, the contact plug 26b and the membrane 10 has an overlap. In Fig. 8a, the overlap of the contact plug 26b is marked with x and the overlap of the membrane 10 with y. The sacrificial layer 22 is present only in the intermediate step shown here and will be removed later.

With appropriate processing, a sacrificial layer 22 of polyimide is deposited on an integrated circuit disk (eg, CMOS technology not shown in the figure). mid applied as a spacer. In the area of the contact plugs 26a, b, the sacrificial layer 22 is opened in the form of a contact hole. In one embodiment, which is shown in FIG. 8a, a metallic contact layer 25 is now deposited and patterned, then a contact metal for the contact plugs 2βa, b is deposited. This metal is etched so that it overlaps an edge of the contact hole. The resistance layer 27 is deposited and patterned. Finally, the sacrificial layer 22 under the membrane 10 is removed, so that it, held by the contact plugs 26a, b, floats above the reflection layer 21 and thus produces a λ / 4 absorber.

Fig. 9 shows a conventional contacting as it is also used in the example of Fig. 8. The contact plug 26b has an overlap x over a diameter z of the contact plug 26b and the membrane 10 has an overlap by a value y on the contact plug 26b.

All embodiments described in US Pat. No. 5,912,464, but also the structures according to US Pat. No. 5,688,699 or from the document http: // www-leti.cea.fr/commun/AR-2003/T5-Photodetection/25 In common, the contact metal protrudes beyond the diameter z of the contact plug 26b (distance x in FIG. 9). The membrane 10 itself protrudes even further (distance y in Fig. 9). The overlaps x and y compensate for adjustment tolerances, they ensure that the area of the contact plug (the contact surface in Fig. 7) is not etched.

FIG. 10 shows how the prior art bolometers scale to reduce pixel size 75. Fig. 10a shows a plan view of the membrane 10 with conventional contacting by the contact plugs 26a and 26b, wherein the membrane 10 is connected via the bridges 76a, b with the contact plugs 26a, b. The bridges 76a, b serve for thermal insulation. As illustrated in FIG. 9, the diaphragm 10 overlaps the contact plug 26b by the value y and the contact plug 26b overlaps the diameter z of the contact plug 26b by the value x. With a reduction (scaling) of the pixel size 75, as shown in Fig. 10b, the size of the contact plug is not scaled due to technology and accordingly the fill factor decreases. One reason for this is that the conventional manufacturing process on photosensitive polyimide is biased as a sacrificial layer 22 and therefore limited to a minimum hole size greater than about 3 μm (further comments below).

FIG. 10a thus shows that, as can also be seen in FIG. 7, the contact plugs 26a, b with their contact with the membrane 10 are relatively large, but their area proportion is still relatively small with a pixel of approximately 50 .mu.m edge length. However, it can already be seen in FIG. 6 that the actual membrane surface 35 has only a relatively small proportion of the total area of the pixel and the fill factor in this embodiment is less than 50%.

As can be seen in FIGS. 6, 8b or 10a, the contact plug 26b is connected to the membrane 10 via a thin arm 20 or 76b. The arm 20 or 76b serves in addition to the mechanical support and the electrical supply in addition to the thermal insulation of the membrane 10 of the contact plug 26b. Its long length and small cross-sectional area provide high thermal resistance between membrane 10 and substrate.

As already described, it is desirable to make the pixels as small as possible. A direct comparison of FIGS. 10a and 10b shows that this is only insufficiently solvable with pixels of conventional technology. In the scaled pixel in Fig. 10b, the contact plugs 26a, b occupy a disproportionate amount of the entire pixel area. This is because the metal of the plug is ne opening through the membrane 10 surmounted by x, also the membrane 10 is typically still larger by y than the overlap x. For a given total area, the proportion of an active area on the membrane 10 becomes smaller, the fill factor decreases, and thus also a sensitivity of the pixels for the IR radiation.

Based on this prior art, the present invention has the object to provide a bolometer and a method for producing a bolometer, wherein the bolometer has no overlap of the membrane 10 or the contact plug and thereby a reduction (scaling) of the pixel dimensions is possible and in which the active area of the membrane 10 is kept as large as possible.

This object is achieved by a bolometer according to claim 1 and a method according to claim 16 or claim 24.

The present invention is based on the finding that it is possible to produce a pixel structure using process steps which are customary, for example, in CMOS technology, which allows significant scaling.

Basic features of a solution can be summarized as an example as follows.

The starting substrate used is, for example, a CMOS wafer which, in the region of the membrane 10 of the bolometer, has a reflector, for example in the form of an Al layer. In the area of the contact plugs or spacers, a respective connection pad (eg of Al) is connected to a read-out circuit. A sacrificial layer, for example, an approximately 2.5 μm thick amorphous silicon layer (a-Si layer), is deposited on this substrate. This can be done, for example, in a CVD method (CVD = c_hemical vapor deposition), possibly assisted by plasma. Subsequently, a first protective layer is deposited (eg a thin layer of silicon oxide in the CVD method, approximately 50-200 nm) so that a layer sequence of the first protective layer / sacrificial layer is formed. Optionally, a stress-compensated layer of, for example, oxide and nitride is deposited instead. The layer sequence is now opened in the area of the spacers. This can be done, for example, by an etching process, wherein in a phototechnics a small contact opening (for example, approximately 0.5 × 0.5 μm ² to 1.5 × 1.5 μm ² ) is exposed in a resist mask. Subsequently, the layer sequence with the resist mask is anisotropically, ie vertically etched, so that a hole extends down to the connection pad (metal connection of the readout circuit). Optionally, the sacrificial layer under the first protective layer can be easily undercut so that the first protective layer overhangs slightly. A thin intermediate layer, for example, of Ti / TiN (eg 20 nm / 80 nm) is sputtered so that a bottom and a perforated wall are at least partially covered. Then, a conductive material is deposited (for example, tungsten in the CVD method) until the hole is completely filled up to a surface. For example, by means of a CMP (CMP = chemical mechanical polishing) method, the conductive material is now polished away from the surface (including the intermediate layer). The hole remains filled with the conductive material. The first protective layer is only polished, but not completely removed.

As a result, there is a basic structure on the basis of which two different continuations of the process are possible.

Process sequence A

A contact layer, for example a thin Ti / TiN layer, is applied to the basic structure and patterned. Then a temperature-sensitive resistive layer (in For example, from a-silicon, optionally also from vanadium oxide (VO _x ) or an organic semiconductor) deposited. The actual measuring resistor of the bolometer is formed by the resistance layer above a narrow gap (gap) in the contact layer. In order to obtain the highest possible thermal insulation of the measuring resistor from the spacers and thus also from the substrate, the gap is preferably arranged as centrally as possible between the spacers in the course of this process control.

By a corresponding configuration of the contact layer, for example by a suitable choice of the layer thickness or layer material, the membrane has a sheet resistance of 377 Ω / D and is therefore suitable as λ / 4 absorber, regardless of the actually higher resistance of the resistance layer.

The resistive layer is now also patterned (for example with lithography and etching step). Next, a second protective layer (covering layer, for example of oxide, optionally of an organic material) is deposited and patterned so that all layers above the sacrificial layer are removed between the membranes, for example in a bolometer array, and between the support arms and the associated membrane are. The resistive layer remains protected all around by the second protective layer and / or an organic overcoat.

Now the sacrificial layer is completely removed through the resulting openings. For this purpose, etching with XeF ₂ , which is described in Chu, PB; JT Chen; R. Yeh; G. Lin; JCP Huang; BA Warneke; KSJ Pister "Controlled Pulse Etching with Xenon Difluoride"; 1997 International Conference on Solid State Sensors and Actuators - TRANSDUCERS '97, Chicago, USA, June 16-19, p. 665-668, which at high rate, isotropic, ie undirected, but with high selectivity, for example, to oxide and organic materials, only the Op- Ferschicht removed. This is particularly effective when the sacrificial layer comprises amorphous silicon. Thus, the membrane is exposed, only supported and contacted by the spacers. The well-protected resistive layer is not etched in this exemplary process. The membrane rests on the spacers. The material of the spacer does not project beyond the resistance layer.

This approach requires only a small number of process steps to realize a bolometer structure. In the following, an alternative process control is described, which has the additional advantage of the largest possible area of the active resistance layer.

Process sequence B

Starting from the same basic structure as before process sequence A, a thin resistance layer (eg of amorphous silicon, VO _x , organic semiconductor) and an insulating layer (for example of oxide) are deposited. Subsequently, these two layers are structured, so that the spacers, which consist for example of tungsten, are exposed. To adapt the sheet resistance of the membrane, a contact layer (for example a thin layer of TiN, 3-15 nm) is applied, optionally followed by a second protective layer, which may have, for example, an oxide. For thermal insulation of the membrane connections to the spacers are now reduced to two narrow webs. This can be done for example by a series of etching steps. In the design of the webs is to make sure that it allows a high fill factor and on the other hand, that the membrane is mechanically stable.

At this time, the contact layer contacts both spacers with a low resistance parallel to actual resistance layer. The contact layer is therefore interrupted in two narrow areas (webs) so that a parallel current conduction through the contact layer is prevented. This can be done, for example, in a further etching step.

The entire structure is now passivated by a thin protective layer (for example, by an oxide layer) to protect the resistive layer. Finally, the sacrificial layer is removed, exposing the membrane. For example, an i-sotropic etching process with XeF ₂ can also be used in this process sequence.

Alternatively, the sacrificial layer can be removed even before a definition of the webs and the insulation of the contact layer. In this case, the resistive layer is protected on all sides against attack of the etching process used by way of example with XeF ₂ without additional passivation.

Significant advantages of a processing according to the invention are thus that the spacers can be scaled to significantly smaller dimensions and still have sufficient adhesion to the membrane 10. Therefore, no openings in the membrane 10 and overlaps x and y are necessary, as was the case with the plugs 26a, b; see Fig. 9.

In other words, the spacers adjoin or end at the bottom of the contact layer without piercing the contact layer.

In addition, the processing according to the invention allows the production of bolometers or bolometer arrays with a significantly smaller pixel size at lower costs.

Preferred embodiments of the present invention will be explained in more detail below with reference to the accompanying drawings. Show it: 1a-i show steps for producing a bolometer according to process sequence A of the present invention and a plan view of the bolometer;

FIGS. 2a-j show steps for producing a bolometer according to process sequence B of the present invention and a plan view of the bolometer;

Fig. 3a-g steps for producing a bolometer with an altered process order; ran and a spacer;

4a shows a plan view of a membrane with contact areas without overlap;

4b is a plan view of a scaled membrane with contact areas without overlap;

Fig. 4c is a cross-sectional view through a portion of a membrane and a spacer;

FIGS. 5a-b are cross-sectional views through conventional microbolometer structures;

Fig. 6 is a perspective view of a conventional structure with membrane;

7 shows a perspective view of a conventional structure with a membrane on two contact plugs with metal overlap;

Fig. 8a is a cross-sectional view through a conventional structure having a contact plug and a portion of a membrane and sacrificial layer still present; Fig. 8b is a plan view of the conventional structure of Fig. 8a;

9 is a cross-sectional view through a Kontaktstöp- and a part of a membrane and marked

overlaps;

Fig. 10a is a plan view of a membrane with conventional

Contacting with contact plug; and

Fig. 10b is a plan view of a scaled membrane with conventional contact with contact plug.

Before the present invention is explained in more detail below with reference to the drawings, it is pointed out that the same elements in the figures are given the same or similar reference numerals and that a repeated description of these elements will be omitted.

1a-h show cross-sectional views of a sequence of steps for a first embodiment of the present invention, and Fig. 1i shows a corresponding plan view with a marked sectional plane 199 of the cross-sectional views.

Figure 1a shows a cross-section of a substrate 100 (e.g., CMOS wafer) on which a pad 110a and a pad 110b are deposited, and also a reflector 120 has been deposited. A connection of the terminal pad 110a and the pad 110b to a underlying CMOS circuit is not shown. Both connection pads 110a, 110b serve for a later contacting of the bolometer.

In the structure shown in FIG. 1 a, a sacrificial layer 130 and a first protective layer 140 are deposited as shown in FIG. 1 b in a subsequent step. The sacrificial layer 130 is removed again in a later step and has a layer thickness, so that the bolometer a represents λ / 4 absorber. In a preferred embodiment, the sacrificial layer 130 comprises amorphous silicon and the first protective layer 140 comprises an oxide.

As shown in FIG. 1c, through holes 150a 'and 150b' are formed through the protective layer 140 and the sacrificial layer 130 in a next step. The passage opening 150a 'is positioned such that it ends on the connection pad 110a, and the passage opening 150b' is analogously positioned so that it ends on the connection pad 110b. In a next step, the passage opening 150a 'and the passage opening 150b' is filled up with a conductive material and excess material is removed, so that a flat surface 142 is formed.

As shown in FIG. 1d, a contact layer 160 is deposited on the surface 142 in a subsequent step. In a next step, which is shown in FIG. 1e, the contact layer 160 is patterned and a resistance layer 170 is deposited. As a result, the structured contact layer 160 has a gap 162 which separates a first part 160a from a second part 160b of the contact layer 160. In order to achieve the best possible thermal insulation of the gap 162 from the spacers 150a and 150b, the minimum distance from the first spacer 150a to the gap 162 should be as close as possible to a minimum distance from the second spacer 150b to the gap 162.

The gap 162 preferably has a width such that the measuring resistance of the bolometer is in a range of, for example, 0.1 kΩ to 1 GΩ and preferably between 1 kΩ and 100 MΩ.

The resistive layer 170 is then patterned and a second protective layer 180 is applied. This is shown in Fig. If. As shown in FIG. In the subsequent step, the surface of the bolometer is structured such that the second protective layer 180 and the contact layer 160 are substantially flush with the spacers 150a and 150b. The patterned resistive layer 170 extends to an interior region of a later-developed membrane surface 192 such that the patterned resistive layer 170 is out of contact with an edge region 190a and 190b. In this step, the first protective layer 140 is also patterned such that the first protective layer 140 is located between the contact layer 160 and the sacrificial layer 130.

In a last step, shown in FIG. 1h, the sacrificial layer 130 is removed. The resulting bolometer has a membrane 10 which has a layer sequence with the first protective layer 140, the contact layer 160 with the first part 160a and the second part 160b, the resistance layer 170 and the second protective layer 180. The bolometer has a surface 192 which is substantially flush with the spacers 150a and 150b. The spacers 150a and 150b have a height 198 which is selected such that the membrane 10 is maintained at a distance 198 and the distance 198 ideally corresponds to one fourth of the wavelength to be detected.

Fig. Ii shows a plan view of the surface 192 of the bolometer with contact surfaces at which the spacers 150a and 150b contact the membrane 10. A dashed line 199 marks the cross-sectional plane which passes through the gap 162 and which is shown in FIGS. 1a to 1h in a viewing direction 81.

Figs. 2a to 2g show a second embodiment of the present invention. 2 a to 2 f show cross-sectional views with reference to a sequence of steps for the manufacture of a bolometer, and FIG. 2 g shows a corresponding plan view with a marked sectional plane 230 the cross-sectional views. The first steps of the second embodiment correspond to a sequence of steps that have been described in Fig. Ia to Ic. A repetition of the explanations for the individual steps is here omitted.

On the structure shown in Fig. Ic, a resistive layer 170 and an insulating layer 210 are first applied, so that the structure shown in Fig. 2a is obtained. FIG. 2 a further shows the substrate 100, the first connection pad 110 a with the first spacer 150 a, the second connection pad 110 b with the second spacer 150 b, the reflector 120, the sacrificial layer 130 and the first protective layer 140.

The resistive layer 170 and the insulating layer 210 are then patterned and the result is shown in FIG. 2b. The structuring takes place in such a way that the resistive layer 170 has no contact with the spacers 150a and 150b, and furthermore the insulating layer 210 does not completely cover the resistive layer 170, so that a first contact point 175a and a second contact point 175b remain free.

As shown in Fig. 2c, a contact layer 160 is applied thereon, which makes contact between the resistive layer 170 and the spacers 150a and 150b.

Subsequently (as shown in FIG. 2 d), the contact layer 160 is first patterned, wherein in particular the contact layer 160 is severed twice by gaps 250 a and 250 b. As a result, the contact layer 160 is divided into a layer 160a having contact with the spacer 150a as well as the resistive layer 170, a layer 160b in contact with the spacer 150b and the resistive layer 170 as well as a layer 160c extending from the layers 160a and the layer 160b is separated. Consequently, the layers 160a and 160b are separated, so that an electric current from the an electric current from the first spacer 150a to the second spacer 150b passes through the resistive layer 170. The layer 160c further has no contact with the resistance layer 170 and has the task of adjusting a sheet resistance of the membrane 10 in accordance with the characteristic impedance of an electromagnetic wave in air.

On the contact layer 160, a second protective layer 180 is then applied. The result is shown in Fig. 2e. Further structuring of the protective layer 180 defines a surface 192 of the membrane 10 of the bolometer.

In a next step, the columns 220a and 220b shown in Fig. 2g are generated. The gaps 220a and 220b sever the membrane 10 with the first protective layer 140, the resistive layer 170, the insulating layer 210, the contact layer 160, and the second protective layer 180. Since a sectional plane associated with the cross-sectional views 2a to 2f does not include the gaps 220a and 220b 220b, columns 220a and 220b are not shown in the cross-sectional views of Figs. 2a to 2f. In the plan view of FIG. 2g, the sectional plane is indicated by the dashed line 230. The arrows 240 show the viewing direction on the cutting plane.

In a final step, shown in FIG. 2f, the first and second protective layers (140, 180) are patterned such that the surface 192 of the membrane 10 terminates substantially flush with the spacers 150a and 150b, and finally the sacrificial layer 130 becomes away.

In a further exemplary embodiment, the structuring of the contact layer 160 takes place asymmetrically, ie the contact layer is only separated by a gap. In this embodiment, the steps are all the way to the structure 2c, identical to the embodiment described above, and a repetition of the description is omitted here.

The structure shown in Fig. 2c is structured in this embodiment as shown in Fig. 2h, i. In particular, only a gap 250 is created, which cuts through the contact layer 160. This results in a layer 160 a, which has contact with the spacer 150 a as well as the resistance layer 170, a layer 160 b, which has contact with the spacer 150 b and the resistance layer 170. Consequently, the layers 160a and 160b are also separated here so that an electric current from the first spacer 150a to the second spacer 150b passes through the resist layer 170. In this embodiment, the sheet resistance of the diaphragm 10 according to the characteristic impedance of an electro-magnetic wave in air can be made by adjusting, for example, the layer 160b or the layer 160a.

The steps shown in FIG. 2i (deposition of the second protective layer 180 and structuring) again correspond to the steps described in FIG. 2e. The same applies to the other steps (generating columns 220a and 220b, further structuring and removing the sacrificial layer 130), which have already been described in the context of FIG. 2f. A renewed repetition is therefore omitted here. Finally, FIG. 2j shows the resulting booter with the membrane 10 and the asymmetrical gap 250.

The specified sequence of steps is only an example and can be changed in further embodiments. Thus, for example, the generation of the gaps 220a and 220b or the formation of the webs 76a and 76b can also take place at the end. The gaps 220a, b are formed such that the largest possible area of the resistive layer 170 is thermally insulated from the spacers 150a, b, and thus the fill factor is as large as possible. At the same time they should provide sufficient support for the membrane 10.

In addition to the process sequence discussed so far, a reversal is also conceivable in which the contact layer 160 is deposited in front of the resistance layer 170. This is shown in Figs. 3a-3g. Again, cross-sectional views are shown, with the first steps again corresponding to a sequence of steps described in FIGS. 1a to 1c. A repetition of the explanations for the individual steps is again omitted here.

The structure shown in FIG. 3 a corresponds to the structure shown in FIG. 1 c and has the first protective layer 140 as the uppermost layer. In this exemplary embodiment, the contact layer 160 is now deposited and structured as the first further layer. The result is shown in Fig. 3b. The structuring takes place in such a way that on the one hand the contact layer 160 is substantially flush with the spacers 150a and 150b and on the other hand has a gap 250 which separates the contact layer 160 into the layer 160a and the layer 160b. The layer 160a is in contact with the spacer 150a and the layer 160b with the spacer 150b.

As shown in Fig. 3c, the insulating layer 210 is deposited thereon and patterned so that the insulating layer substantially fills the gap 250 and also exposes the first pad 175a to the layer 160a and the second pad 175b to the layer 160b.

As shown in Fig. 3d, the resistive layer 170 is deposited thereon and patterned so that the resistive layer 170 terminates substantially flush with the standoffs 150a and 150b.

As shown in FIG. 3e, the second protective layer 180 is again applied and patterned on it, so that the Membrane 10 is defined with the surface 192. The result is shown in Fig. 3f. As a last step, in turn, the sacrificial layer 130 is removed, so that the structure of FIG. 3g is formed.

FIG. 4 a shows a plan view of the membrane 10 with contact surfaces, where the spacers 150 a and 150 b contact the membrane 10.

Fig. 4b shows the scaled membrane 10, i. a correspondingly reduced membrane 10. Here, in contrast to the prior art, the contact surfaces 150a and 150b also corresponding to a size of the membrane 10. In both cases, the membrane 10 shows no overlap over contact surfaces where the spacers 150a and 150b, the membrane 10 contact.

Fig. 4c shows a scaled contact between the membrane 10 and the spacer 150b. The membrane 10 lies without overlapping on the spacer 150b.

A method according to the invention is advantageous in several respects compared with the prior art. Thus, a process according to the invention with the spacers 150a and 150b, which preferably have tungsten, and with the sacrificial layer 130, which preferably comprises amorphous silicon (a-Si), permits a reduction of the IR-sensitive pixel. A conventional process with photosensitive polyimide has a minimum hole size that must be greater than about 3 microns. Even if smaller holes in the polyimide would be possible (eg by a multi-layer mask of photoresist and oxide on the polyimide, which can then be opened with an anisotropic etching process with oxygen plasma), these can not or insufficiently filled, for example, with tungsten. The tungsten deposition in the CVD process typically requires temperatures above 45O ⁰ C, at which the polyimide is no longer stable. On the other hand, use of a-Si as the sacrificial layer 130 is temperature-resistant and allows spacers 150a and 150b to be deposited, for example, from tungsten in good quality, as is customary in CMOS technology in multilayer metallization. For example, holes of very small diameter and high aspect ratio (depth / diameter) can be etched in the a-Si layer, as known from the production of trenches in DRAMs. The a-Si layer is stable, so that a relatively strong etching-back process, eg with Ar ions, is possible before the deposition of the contact layer 160 (for example by sputtering of Ti / TiN). This lowers a contact resistance between the spacers 150a, b and the contact layer 160 and improves the adhesion of the contact layer 160 to the spacer 150a, b.

The resulting structure with membrane 10 overlying spacers 150a, b can be scaled to a small size because the stated process steps (except for the deposition and isotropic removal of exemplary a-Si sacrificial layer 130) can be taken from an advanced CMOS process. Thus, a 0.25 .mu.m process allows a diameter for the spacers 150a, b which is smaller than 0.5 .mu.m, the holding arms can be as wide as a diameter of the spacers 150a, b and have a distance of 0.25 .mu.m from the membrane 10.

Substantial advantages of a processing according to the invention are thus that the spacers 150 a, b can be scaled to significantly smaller dimensions and still have sufficient adhesion to the membrane 10. Therefore, it is not necessary to pass the spacers 150a, b through the membrane 10 and overlap by the values x and y, as was the case with the plugs 26a, b. In an exemplary embodiment of the present invention, it is furthermore possible to dispense with the formation of webs 76a, b, which leads to a further increase in the filling factor and to improved mechanical stability. In addition, the processing according to the invention allows the production of bolometers or bolometer arrays with a significantly smaller pixel size at lower costs.

This allows pixels with 20 x 20 μm ² or 15 x 15 μm ² with a constant high filling factor. The distance between the membranes 10 in a bolometer array can be 0.5 μm, for example, so that a pixel pitch (distance from center of the pixel to the center of the pixel) can also be 15-20 μm.

As stated above, two embodiments of the present invention are based on two process flows. By specifying preferred materials, layer thicknesses, methods used, etc., both process sequences can be summarized as follows.

Process flow A

- Providing a CMOS disc with passivated surface

Depositing a metallic reflector 120 and two connection pads 110a, 110b for a compound CMOS membrane, e.g. made of thin Al (e.g., 100-200 nm, therefore only small)

Deposition of a-Si about 2.5 μm (as sacrificial layer 130)

- possibly smoothing the surface by a CMP process

- Oxide deposition of a first protective layer 140 (approximately 200 nm) - Define through holes 150a 'and 150b' by means of photo technology (diameter about 0.5-1 microns)

Oxide etching, silicon etching anisotropic, stop on pad metal from pads 110a and 110b

Ti / TiN barrier sputtering in the through holes 150a 'and 150b'

Tungsten CVD method for filling through holes 150a 'and 150b' CMP process for removing the tungsten and Ti / TiN from the surface

- sputtering back

Sputtering the contact layer 160; TiN thin (for layer resistivity of 377 Ω / D)

Etching the contact layer 160 by means of photo technology (removing TiN below the actual resistance), forming a gap 162

Deposition of a-Si doped for bolometer resistance 170 -photo technique, etching a-Si, patterning of resistive layer 170

Deposition of oxide in a CVD process to form second protective layer 180 (about 30 nm)

- Photographic technique for defining the membrane surface and exposing the connecting arms

Etching the layers into the a-Si sacrificial layer 130

Removal of the a-Si sacrificial layer 130, eg with highly selective (oxide is hardly attacked), isotropic etching in gaseous XeF ₂

Process flow B

- Providing a CMOS disc with passivated surface

Depositing a metallic reflector 120 and two connection pads 110a and 110b for a compound CMOS membrane, e.g. made of thin Al (e.g., 100-200 nm, therefore only small)

Deposition of a-Si about 2.5 μm (as sacrificial layer 130)

- possibly smoothing the surface by a CMP process

Oxide etching, silicon etching anisotropic, stop on pad metal from pads 110a and 110b Ti / TiN barrier sputtering in the through holes 150a 'and 150b'

Tungsten-CVD method of filling through holes 150a 'and 150b' - CMP for removing the tungsten and Ti / TiN from the surface

- sputtering back

Deposition a-Si doped for bolometer resistance 170

Deposition of oxide in a CVD process to form the insulating layer 210 (about 30 nm)

Photographic technique for patterning the oxide of the insulating layer 210

- Photographic technique for structuring the bolometer resistor 170 - Sputtering of the contact layer 160; TiN thin (for sheet resistance of 377 Ω / D)

Deposition of oxide in a CVD process as second protective layer 180 (about 30 nm)

- Photographic technique for defining the narrow ridge areas 76a and 76b, etching of oxide (second protection layer 180), TiN

(Contact layer 160), a-Si (resistance layer 170) and again oxide (first protection layer 140).

Photographic technique for insulating the TiN layer 160 of the membrane 10 - etching oxide of the second protective layer 180 and TiN of the contact layer 160 and generating the gaps 250a and 250b

Deposition of oxide in a CVD process for protecting the contact layer 160 (about 30 nm)

Removal of the a-Si sacrificial layer 130, for example with highly selective (oxide is hardly attacked), isotropic etching in gaseous XeF ₂

The above materials are only examples that allow very good process control. For example, some alternatives include the following substitutions. The sacrificial layer 130 of a-Si can be etched with an alternative ClF ₃ (Clorfluorid) or with an isotropic SF ₆ plasma (pivoting ^¬ felfluorid plasma). The sacrificial layer 130 may also include a temperature resistant polymer (eg, polyimide). The through holes 150a 'and 150b' for the spacers 150a and 150b can then be etched with anisotropic O ₂ plasma, the sacrificial layer 130 can then also be removed by a 0 ₂ ~ plasma.

When the sacrificial layer 130 is removed in an etching step, it is important to protect the resistive layer 170 and / or the contact layer 160 during the etching step. For this, the presence of the protective layer 140 is advantageous. In this case, the material is preferably selected such that it is not or hardly attacked during the step of removing the sacrificial layer 130. However, if a method is available which removes the sacrificial layer 130 without attacking the resistive layer 170 and / or the contact layer 160, the first protective layer 140 may be dispensed with in another embodiment.

The temperature-dependent resistance layer 170 may comprise, for example, another semiconductor material (VO _x , GaAs, organic semiconductor or the like). Instead of the silicon oxide layers, it is also possible to use layers of silicon nitride (or a combination of both).

Claims

claims

1. Bolometer with the following features:

a membrane (10);

a first spacer (150a); and

a second spacer (150b),

wherein the membrane (10) comprises a resistive layer (170) and a contact layer (160), wherein the contact layer (160) has a first contact region on a background-facing side, where the first spacer (150a) electrically contacts the contact layer (160) and having a second contact region at which the second spacer (150b) electrically contacts the contact layer (160) and the first and second spacers (150a), (150b) at a predetermined distance (198) toward the membrane (10) keep the ground, and

wherein the contact layer (160) is interrupted laterally by a gap (162) so that the contact layer (160) is divided into at least two parts, of which the first part (160a) has the first contact area and the second part (160b). has the second contact region and within the contact layer (160) no direct connection of the first contact region and the second contact region exists, and

wherein the resistance layer (170) is in contact with the first part (160a) of the contact layer (160) and the second part (160b) of the contact layer (160).

2. A bolometer according to claim 1, wherein the membrane (10) further a first protective layer (140); and

a second protective layer (180),

wherein the resistive layer (170) and the contact layer (160) extend between the first protective layer (140) and the second protective layer (180).

The bolometer of claim 1 or claim 2, wherein the membrane (10) terminates laterally substantially flush with the first spacer (150a) and the second spacer (150b).

A bolometer according to any one of claims 1 to 3, wherein the predetermined distance (198) is equal to one quarter of an infrared wavelength.

5. Bolometer according to one of claims 1 to 4, wherein the contact layer (160) has a layer thickness, so that a sheet resistance of the membrane (10) corresponds to the characteristic impedance of electromagnetic waves in air.

6. Bolometer according to one of claims 1 to 5, wherein the substrate has a reflector (120).

7. A bolometer according to any one of claims 1 to 6, wherein the membrane (10) has a substantially quadrangular shape in plan view from a side facing away from the ground, and the first spacer (150a) and the second spacer (150b) abut each other located opposite corners.

8. The bolometer according to any one of claims 1 to I ₁ wherein the gap (162) has a width such that the resistance contribution to a measured resistance of the Bolome- ters by a current flow between the first con- and the second contact region is approximately between 1 kΩ and 100 MΩ.

9. The bolometer of claim 1, wherein a position of the gap is selected so that an electric current from the first spacer to the second spacer causes the gap to pass reaches a distance and the distance is as equal as possible to a distance covered by the electric current from the gap (162) to the second spacer (150b).

10. A bolometer according to any one of claims 1 to 9, wherein the membrane (10) further comprises an insulating layer (210) disposed between the resistive layer (170) and the contact layer (160) and the resistive layer (170) the background facing the ground is located.

The bolometer according to any one of claims 1 to 9, wherein the diaphragm (10) further comprises an insulating layer (210) disposed between the contact layer (160) and the resistance layer (170) and the contact layer (160) is disposed on the Background facing side.

The bolometer of any one of claims 1 to 11, wherein the membrane (10) further comprises a first land (76a) and a second land (76b), and the first land (76a) comprises the membrane (10) with the first spacer

(150a) and the second web (76b) connects the membrane

(10) connects to the second spacer (150b) so that the membrane (10) from the first spacer

(150a) and the second spacer (150b) is thermally insulated.

13. A bolometer according to any one of claims 1 to 12, wherein the resistance layer (170) comprises amorphous silicon or vanadium oxide.

14. A bolometer according to any one of claims 1 to 13, wherein the contact layer (160) comprises titanium nitride.

A bolometer according to any one of claims 1 to 14, wherein the first spacer (150a) and / or the second spacer (150b) comprise tungsten or copper.

16. The bolometer of claim 1, wherein the substrate further comprises a first terminal pad and a second terminal pad for contacting the first spacer and the second spacer.

17. A method of making a bolometer comprising the steps of:

a) providing a substrate (100);

b) applying a sacrificial layer (130) to the substrate (100);

c) forming a first passage opening (15Oa ') and a second passage opening (15Ob');

d) forming first and second spacers (150a), (150b) in the first and second through holes (15Oa '), (15Ob'), respectively;

e) applying a contact layer (160) such that the contact layer (160) on a side facing the substrate (100) has a first contact region where it is contacted by the first spacer (150a) and has a second contact region , to the the same is contacted by the second spacer (150b);

f) patterning the contact layer (160) to form a gap (162) therein, such that the contact layer (160) is divided into two parts, of which the first part (160a) has the first contact area and the second part (160b ) has the second contact region and within the contact layer (160) there is no direct connection from the first contact region to the second contact region;

g) applying a resistive layer (170) such that the resistive layer (170) is in contact with the first part (160a) of the contact layer (160) and the second part (160b) of the contact layer (160), the resistive layer (170) 170) and the contact layer (160) form a membrane (10) of the bolometer;

h) structuring an outline of the membrane (10); and

i) removing the sacrificial layer (130).

A method according to claim 17, further comprising a step of depositing a reflector (120) between step a) and step b).

A method according to claim 17 or claim 18, further comprising, between step f) and step g), a step of applying an insulating layer (210) such that the insulating layer (210) covers the first part (160a) at a first contact point (175a) and the second part (160b) at a second contact point (175b) leaves free.

20. The method of claim 17, wherein the step of applying a sacrificial layer is performed such that the sacrificial layer has a layer thickness that corresponds to a quarter of an infrared wavelength.

21. The method according to any one of claims 17 to 20, wherein the step of applying a contact layer (160) is such that the contact layer (160) has a layer thickness, so that the sheet resistance of the membrane (10) a characteristic impedance of an electromagnetic wave in air equivalent.

22. The method according to any one of claims 17 to 21, wherein the step a) comprises the sub-steps:

al) providing a substrate (100); and

a2) forming a first connection pad (110a) and a second connection pad (110b),

wherein the first one of the first terminal pad (110a) contacts the first spacer (150a) and the second terminal pad (HOb) contacts the second spacer (150b).

23. The method of claim 17, wherein the step of depositing a sacrificial layer comprises using amorphous silica, or vanadium oxide.

24. A method according to any one of claims 17 to 23, wherein the step of applying a contact layer (160) comprises using titanium nitride and the step of filling the first and second through holes (15Oa '), (15Ob') comprises using tungsten or copper.

25. The method according to any one of claims 17 to 24, wherein the step b) comprises a chemical vapor deposition.

26. A method of making a bolometer comprising the steps of:

a) providing a substrate (100);

b) applying a sacrificial layer (130) to the substrate (100);

d) forming a first spacer (150a) in the first passage opening (15Oa ') and a second spacer (150b) in the second passage opening (15Ob');

e) applying a resistive layer (170) laterally such that the resistive layer (170) is not in contact with the conductive material in the first and second through holes (15Oa '), (15Ob');

f) applying an insulating layer (210) on a side of the resistance layer (170) facing away from the substrate (100) so that the insulating layer (210) forms the resistance layer (170) at a first contact point (175a) and at a second contact point (175b). leaves free;

g) applying a contact layer (160) such that the contact layer (160) on a side facing the substrate (100) has a first contact region, where it is contacted by the first spacer (150a), and a second contact region where it is contacted by the second spacer (150b), and the contact layer (160) contacts the first contact pad (175a) and the second contact pad (175b) of the resistive layer (170);

h) structuring an outline of the membrane (10);

i) patterning the contact layer (160) to form at least one gap (250a) therein, such that the contact layer (160) is divided into at least two parts, of which the first part (160a) is the first contact area and the first contact point (175a) of the resistance layer

(170) and the second part (160b) has the second contact region and the second contact point (175b) of the resistance layer (170) and there is no direct connection between the first contact region and the second contact region within the contact layer (160);

j) removing the sacrificial layer (130).

27. The method according to claim 26, further comprising the step of:

Patterning the membrane (10) to form therein a first land (76a) and a second land (76b), the first land (76a) contacting the first spacer (150a) and the second land (76b) has contact with the second spacer (150b), and the first land (76a) and the second land (76b) thermally insulate as much of the resistive layer (170) from the first spacer (150a) and the second spacer (150b) ,

28. A method according to claim 26 or claim 27, further comprising the following steps between step a) and step b):

Depositing a reflector (120);

Forming a first connection pad (110a) and a second connection pad (110b); and

wherein the first terminal pad (110a) contacts the first spacer (150a) and the second terminal pad (110b) contacts the second spacer (150b).

29. The method of claim 26, wherein the step of depositing a sacrificial layer is performed a layer thickness such that the sacrificial layer has a layer thickness that is one quarter of an infrared wavelength.

30. A method according to any one of claims 26 to 29, further comprising a step of depositing a first protective layer (140) and / or depositing a second protective layer (180), wherein the step of depositing the first protective layer (140 ) between step b) and step c) and the step of depositing the second protective layer (180) between step i) and step j).

31. The method of claim 26, wherein the step of patterning the contact layer comprises forming a second gap such that the contact layer further comprises a third portion, and wherein the Step of applying a contact layer (160) is carried out such that the third part (160c) has a layer thickness, so that a sheet resistance of the membrane (10) corresponds to a characteristic impedance of an electromagnetic wave in air.